Clastic injection: process to product

(1)

Clastic injection:

process to product

Jessica Amy Ross

Submitted in accordance with the requirements for the degree of

Doctor of Philosophy

The University of Leeds

School of Earth and Environment

(2)

Declaration of authorship

The candidate confirms that the work submitted is her own, except where work which has formed part of jointly-authored publications has been included. The contribution of the candidate and the other authors to this work has been explicitly indicated below. The candidate confirms that appropriate credit has been given within the thesis where reference has been made to the work of others.

The work in Chapter 5 reproduces a manuscript that was published in Sedimentology in 2011.

Ross, J.A., Peakall, J. and Keevil, G.M. (2011) An integrated model of extrusive

sand injectites in cohesionless sediments. Sedimentology, 58, 1693-1715.

Data were collected in the laboratory by Jessica Ross. All data were processed, interpreted, presented and the conceptual model designed by Jessica Ross. Ideas were shaped during discussion with co-authors.

The work in Chapter 7 reproduces a manuscript that was published in the Journal of the Geological Society in 2013.

Ross, J.A., Peakall, J. and Keevil, G.M. (2013) Sub-aqueous sand extrusion

dynamics, Journal of the Geological Society, 170, 593-602.

Fieldwork was completed by Jessica Ross. All presented data were collected and interpreted and the model designed by Jessica Ross. Ideas were shaped during discussion with co-authors.

(3)

Chapter 8 reproduces a manuscript that has been accepted for publication in

Sedimentology.

Ross, J.A., Peakall, J. and Keevil, G.M. DOI: 10.1111/sed.12115. Facies and fluid

flow of sandstone-hosted columnar intrusion: the pipes of Kodachrome Basin State Park.

Fieldwork was completed by Jessica Ross. All presented data were collected, analysed and interpreted by Jessica Ross. Ideas were shaped and developed through discussion with co-authors.

This copy has been supplied on the understanding that it is copyright material and that no quotation from the thesis may be published without proper acknowledgement.

The right of Jessica Amy Ross to be identified as Author of this work has been asserted by her in accordance with the Copyright, Designs and Patents Act 1988.

(4)

Acknowledgments

My most sincere thanks go to my supervisors Jeff Peakall and Gareth Keevil who have unfailingly supported me and provided invaluable guidance from the inception of the research project, through to the final thesis. I would also like to thank Russell Dixon for his technical support during the experimental phases of this project and Eric Condcliffe for his exhaustive knowledge of all things SEM. Gareth Keevil and Rachael Dale provided important support in the field in Ireland and the USA respectively. Mads Huuse acted as external examiner for this thesis who, along with Dave Hodgson is thanked for a rigorous and enjoyable viva. Also, for providing comments which have challenged me to improve this thesis.

There are some people without whom I would, quite literally have gone insane, firstly Rich who has been my constant anchor to reality through his love and unfailing support. I have been lucky enough to share office 7.131 with fellow troglodytes; Alan, Hollie, Jo, Rachael, Sarah and Steve who have managed to make me laugh when I never thought I could. I couldn’t have done it without you. Finally, my parents, brother and sisters never allowed me to believe that I couldn’t do it, so I got on and did it.

Funding for this research has come from the Natural Environment Research Council PhD studentship number NE/H524673/1 with additional financial support gratefully received from the Earth Surface Science Institute at the University of Leeds and the 2011 Gill Harwood Memorial Award from the British Sedimentological Research Group.

(5)

Abstract

Subsurface sediment remobilisation and subsequent extrusion records the release of overpressure through a sealing lithology by an injecting slurry. This investigation focuses upon injections occurring in the shallow-subsurface and utilises a multidisciplinary approach to reappraise the dynamics of sand injections across a variety of scales. The study provides detailed analysis from the laboratory and field on three scale of form: centimetre-scale fluidisation pipes, metre-scale extrusions and decimetre-scale fluidisation pipes. The investigation has helped bridge the knowledge-gap and converge ideas between traditional geologically-derived interpretations of sand injections and fluidisation pipes considered in chemical engineering. Laboratory modelling of fluidisation pipes provides the first process-based model of shallow sub-aqueous sand injection and extrusion in cohesionless sediments and recognises a series of processes hitherto unlinked to previously described internal sedimentary structures in fluidisation pipes. Fluidisation is shown to occur through a series of discrete phases and critically, the style, stability and temporal evolution of piping, along with flow velocity and concentration, are shown to exhibit considerable variability. The novel application of particle tracking velocimetry to active sand injections suggests that this technique could be invaluable in unravelling the flow dynamics in active injections. A process-based mechanism of sand extrusion formation is proposed though investigation of the internal architecture of seismically-induced sub-aqueous sandstone extrusions. Sand sheets are shown to form through deposition from gravity currents when multiple vents extrude coevally, whereas sand mounds or volcanoes will form from a single vent unless bypassing mechanisms such as channelisation influence sediment deposition. Previous estimate of flow velocity in sandstone intrusions and Reynolds numbers are shown to be inaccurate by up to two orders of magnitude. The investigation also demonstrates that sandstone-hosted intrusions exert control on basinal fluid flow in a manner previously identified only in mudstone-hosted intrusions and proposes a new model of the formation of the sandstone intrusions in Kodachrome Basin State Park.

(6)

Thesis Structure

Title: ... i Declaration of authorship... ii Acknowledgments ... iv Abstract ... v Thesis Structure ... vi Table of Contents ... vi

List of Figures ... xii

List of Tables ... xxi

Nomenclature... xxii

List of Figures

Figure 2.2. Schematic block diagram showing the elements of a sandstone injection

complex. Parent sand body geometry is marked by dashed line. Modified from Hurst et al. (2011). ... 6

Figure 2.3. Conical sandstone intrusions from the Faroe-Shetland basin showing

characteristic discordance. Arrows denote the tops of the intrusions. From Cartwright et al. (2008). ... 8

Figure 3.1. Compressibility of mixed particle systems; (a) as deposited; (b) effects of

compression and shearing. Redrawn from Yamamuro and Lade (1999). ... 13

Figure 4.1. Basic set-up for a fluidised bed. (1) water inlet, (2) inlet section, (3)

homogenising section, (4) distributor plate, (5) column containing particulate bed. Adapted from Asif et al. (1992). ... 22

Figure 4.2. Determination of the minimum fluidisation velocity from the pressure

drop across a bed (left) and the bed height (right). A-B region is where particles are in a fixed bed, at the base of a column. Fluidisation begins at point B, flow rate decreases from point C. Incipient fluidisation is point D. Redrawn from Couderc (1985). ... 22

Figure 4.3. Richardson-Zaki plot of dimensionless fluidisation velocity against

time-averaged mean voidage to demonstrate discontinuous bed expansion. From Didwania and Homsy (1981). ... 26

Figure 4.4. Classification of fluidisation behaviour of solids (horizontal axis) by

ambient water (left, vertical dimensional axis) or by any fluid (right, vertical dimensional axis). pp is particle pressure, dp grain diameter and and De is a

dimensionless density number (Redrawn from Di Felice (1995). ... 28

Figure 4.5. (A) The disparity in slope between experimental settling velocities

(symbols) and Richardson-Zaki predictions (lines). Whereas when wall effects are ignored (B), experimental values of n as a function of Re (symbols) are in good agreement with Richardson-Zaki predictions (line). Adapted from Di Felice and Kehlenbeck (2000). ... 32

Figure 4.6. Bed inversion phenomenon. Black filled circles represent type 1

(13)

particles. From Di Felice (1995), based on experimental observations of Moritomi et al. (1986). ... 34

Figure 5.1. A) Schematic illustration of experimental set up, B) Details of inlet pipe

configurations. (Run 6 middle configuration, Run 7, bottom configuration, all others as top). ... 42

Figure 5.2. A series of video stills demonstrating the most significant periods in the

structural evolution of a run (here Run 4). (A) Starting 3-layer stratigraphy. Apparent slight doming on the upper surface of the sediment and in the middle layer is due to barrel distortion. (B) Initiation of void. (C) Rupture occurs and pipe propagates to surface. Infiltration horizon can be seen above the dish-shaped fine layer following fluidisation. (D) Venting at surface and initiation of secondary pipe. (E) Piping has stabilised and fine material can be seen lining the inner edges of the pipes. (F) Final geometry showing pipes and sand volcanoes. Much of the fine layer has been elutriated and redeposited on top of the sediment column. The times at which these stills were taken are as follows (in minutes and seconds): A – 0.00, B - 2.22, C – 8.30, D – 8.33, E – 9.43, F – 11.05. Scales shown are in centimetres (lower) and inches (upper). ... 46

Figure 5.3. (A) Diagram of the inclined pipe in Run 4 demonstrating stress pipes in

the infiltration horizon above the pipe. Scale bar 1 cm. (B) Red arrows highlight stress pipes aligned at a high angle to the walls of a clastic sill. Coin for scale is 3 cm diameter (adapted from Diggs (2007)). ... 49

Figure 5.4. Preservation of initial rupture zone results in fine sediment swirls

preserved in surrounding coarse sediment. From Run 2, at 0.42 minutes (upper) and 1.57 minutes (lower). Key features have been outlined in B to enhance clarity. Scales shown are in centimetres (lower) and inches (upper). ... 51

Figure 5.5. Pipe morphologies from 2 runs. Video frames have been interpreted for

clarity. A) From Run 5 at 0.59 minutes. The main pipe is migrating to the left whilst the pipe to the right stays stationary, both maintaining a vertical morphology. A poorly defined zone of fluidised, convecting sediment exists to the right of the main pipe. Sand ball morphology is seen to the right of the right-hand pipe. B) From Run 7 at 1.52 minutes. There are eight, clearly defined pipes venting water and fluidised

(14)

sediment, each separated by unfluidised sediment from the original top layer or clasts of incorporated top-layer material. Black dots are 1 cm spaced. ... 53 Figure 5.6. Series of stills from Run 10 demonstrating the development of dyke and sill style piping. A) Following rupture and piping onset, a new pipe begins to migrate upwards and horizontally from a deformed sediment boundary. B) Horizontal migration is abandoned in favour of vertical migration. C) Fluidised sediment “ponds” and migrates laterally beneath an impermeable heterogeneity, either pre-existing in the upper sediment layer or due to fine sediment amalgamating on the upper edge of the sill. D) The pipe reaches the sediment surface and begins venting sediment. This morphology remains stable for the duration of this run. Black dots are 1 cm spaced. ... 56

Figure 5.7. Effects of pipe migration on layering. The pipe has migrated to the left

from out of view on the right of the photo. The original fine ballotini layer has been completely removed and the boundary between the coarse ballotini and overlying silica sand is difficult to discern due to fluidisation. Aggregates of fine ballotini that were originally in the middle layer are left suspended in the silica sand layer, which also contains coarse ballotini. This is evident from the colour difference on the right and left of the pipe. From Run 9 at 6.23 minutes. Black dots are 1 cm spaced. ... 58

Figure 5.8. Photograph of an excavated sand volcano orientated with the sediment

surface at the top. The topographic feature has relief of 4 mm and has a clear vent. Dashed lines denote the region where a pipe would be expected if it were present. Scale in mm. ... 60

Figure 5.9. Topographic relief created through venting. Important features are

labelled and topography marked by the dashed lines. A) Run 4 at 10.22 minutes. The sand volcano is approximately 2 cm high and has two vents fed by separate pipes. The pipe to the right is venting sediment at a faster rate than the left hand pipe. Horizontal scale is 30 cm. B) Run 11 at 7.30 minutes. Pipe is migrating to the left of the picture and relict topography of oviform sand mounds is visible to the right of the picture. C) Run 7 at 5.30 minutes. Multiple pipes have produced an undulating topography, with an extinct pipe (centre, marked by thick arrow) in the depression between two active sand volcanoes and an active vent on the far right

(15)

with negative topographic relief. Black dots are 1 cm spaced. D) Run 4 post-experiment vent, millimetre scale visible. E) Run 9 at 1.47 minutes, post-post-experiment. Fine sediment can be seen draped over an asymmetric vent. To the right of the vent sediment flows comprising the sand volcano flanks are clearly visible. F) Internal structure of sand volcano in Run 4. ... 61

Figure 5.10. Schematic representation of a new model of pipe formation in

cohesionless sediments. A) Starting conditions with a fine sediment layer between two, thick, coarse sediment layers. When overpressure is generated heterogeneities in the upper layer determine whether the system develops into B) (homogeneous upper layer) An even void forms beneath the fine sediment layer immediately followed by infiltration of fine sediment into the top layer and stress pipe formation (partial fluidisation) or C) (heterogeneous upper layer) An uneven void is formed with an inverted conical shape. Infiltration into the upper layer occurs with localised absence of infiltration horizon at void apex. From these initial stages the following occurs: B1) Rupture and initial pipe formation. Failed rupture zones may be preserved and a fluidised zone develops at the base of the pipe which is now elutriating fluidised sediment. This configuration is not stable. C1) In this scenario, one or two rupture zones may have occurred resulting in two pipes elutriating sediment and water. Most of the fine sediment has been removed, although a small amount remains between pipes. C2) here multiple ruptures have occurred, leading to multiple pipes. Most fine sediment has been elutriated and a zone of fluidised sediment exists below the base of pipes. An unstable configuration which will lead to one of the pipes becoming dominant and others shutting down. D1) A dominant pipe has stabilised and the rim is lined with fine sediment; preventing pipe migration. A localised zone of fluidised sediment exists around the vent with down bending of all layers orientated towards the base of the pipe. D2) Unconfined pipe migrates laterally. Original stratification is undisturbed on the leading side of the pipe whereas where the pipe has passed, sediment is homogenous and relict topography is preserved on the surface. A zone of circulating fluidised sediment is associated with the rear of the migrating pipe. E) Schematic demonstrating gravity currents flowing down the sides of a sand volcano to develop sand sheets and

(16)

fine-grained sediment is held in suspension before settling and draping sand volcano. This model does not take into account atypical features such as double void apex formation, dyke and sill intrusions, unusual topography and infiltration associated with sub-vertical pipes. ... 76

Figure 6.1. Diagram showing the range of outcrop evidence for low regime in

injections. Heavy arrows denote flow direction in each example. Flow in G is into the page. ... 82

Figure 6.2. Velocity as a function of sand fraction; r is conduit radius. From (Gallo

and Woods (2004). ... 85

Figure 6.3. Velocity (u) as a function of conduit radius (r) for a given overpressure at

the base of the fracture (s) is sand fraction. From Gallo and Woods (2004). ... 86

Figure 6.4. Schematic showing the experimental set up. Dotted area is the gravel

and mesh baffle. All input pipes are shown in blue. Those from the manifold to the test tank are all the same length (1 m). ... 92

Figure 6.5. Photograph of the upper sediment layers to demonstrate stratigraphy.

Apparent curve of fine layer is due to barrel distortion arising from tank curvature. The fine Ballotini layer is 1 cm thick. ... 92

Figure 6.6. (A,B) Frames extracted from video files to demonstrate the very

fine-detail achievable using a commercially available digital video camera (Nikon 3100) when the particles are stationary (right of image) compared to streaking and lack of detail when particles are in motion (left of image). (C) Sample frame from the Phantom camera. Flow in the pipe is from the base left to the top right; and black “seeding” particles are visible in the centre of the pipe. ... 93

Figure 6.7. Schematic to show particle tracking between frame-pairs. ... 95 Figure 6.8. Photographs to demonstrate how the high speed camera and LED lights

are mounted relative to the tank. A – Led lights, B – Tank wall. ... 96 Following a successful experimental run in which a fluidisation pipe passed through the field of view, the following steps were undertaken to process the images ready for particle tracking: (i) isolation of the area of interest in the image sequence using image masking in order to minimise noise; (ii) application of a Sobel operator to increase individual particle visibility (see details below); (iii) calculation

(17)

of the arithmetic mean of all 100 images in the stack by computing the average values of the pixels’ intensity in pre-defined interrogation areas; (iv) subtraction of the calculated image mean from each subsequent image in the stack to remove any stationary features in the image; (v) subtraction of a duplicate of the first image from the stack (this further reduces/removes stationary features); (vi) subtraction of a duplicate of the last image from the stack (this further reduces/removes stationary features); and, (vii) conversion to double images as required for the particle tracking steps outlined in section 6.3.1. Note that the Sobel operator is an edge detection filter that performs a 2-D spatial gradient measurement on an image and thus highlights the particle edges, having the overall effect of increasing the contrast of the seeding particles. ... 96

Figure 6.10. Close-up of the infiltration horizon that commonly develops above the

fine-grained layer. ... 100

Figure 6.9. Pressure trends from runs 1-5. Left hand images are the pressure trends

for the complete duration of each experimental run with the exception of the end of run 4 (D) and the beginning of run 5 (E). The vertical grey bars highlight the time period shown in the right-hand graph (‘). Red arrows 1 and 2 denote consecutive rupture events. A 20-point moving average has been applied to the 20 s intervals (black line) to remove noise and improve visualisation of the pressure trends. ... 101

Figure 6.11. Inclined infiltration of fine-grained sediment prior to the arrival of a

migrating fluidisation pipe (light coloured zone to left of images). Secondary grain (white arrow) shows sub-horizontal path. Initial and final grain locations denoted by black circles. Total time elapsed is 1 second. ... 102 Manually measured flow velocities of particles at various points across the pipe diameter are presented below (Table 6.2). These measurements capture individual particles that the code is, as yet, unable to track. All particles were tracked over a 100ms period to ensure that the same particle remained in field of view. ... 105

Figure 6.12. From experiment 5. Time-series of interpreted video frames over a 20 s

period (every 10th_{frame) demonstrating particle trajectories and the variety of}

morphologies at any given time. Frames run from top left to bottom right. Frames are 16 mm wide. ... 106

(18)

Figure 6.13. (Left) From Experiment 4, showing every-other frame extracted over 2

seconds of video (left) to demonstrate the spatial and temporal variation in pipe morphology. (Right) Interpretation of the video frames showing particle trajectories (arrows) and flow zones. The pipe is migrating from left to right. Frames are 16 mm wide. Key is the same as Figure 6.12. ... 107

Figure 6.14. The horizontal and vertical components of the flow across the whole

field of view, as derived by the PTV code. Red box denotes area shown in Figure 6.15. ... 108

Figure 6.15. Horizontal and vertical velocity vectors below 1 x 10-2_ms-1_{. ... 109}

Figure. 6.17. Series of consecutive images demonstrating the trajectory of a particle

suspended in the main flow over 40 milliseconds. Image is 23.57 x 17.33 mm. ... 112

Figure 6.18. The pressure field at the pipe wall associated with travelling waves. m0₌

number of eigenvalues in the plane. Dark red corresponds to negative values, white to positive values. From Wedin and Kerswell (2004). ... 119

Figure 6.19. Axial section of a pipe at different downstream positions showing

three-fold azimuthal symmetry in a travelling wave at Re = 1250. High speed streaks in the downstream flow component at the pipe wall (red) are more stable than low speed streaks (blue) near the centre of the pipe. Velocity components are shown by arrows, negative values are red and positive values are blue. From Faisst and Eckhardt (2003). ... 119

Figure 7.1. Location map of the study area. ... 135 Figure 7.2. Photomontage of the outcrop at Freagh Point, facing West. Faults are

marked and sand volcano vents are shown by white dots along with facies distribution in the underlying interpretation. Yellow 1 m scale bar to centre-left. See text for facies codes ... 138

Figure 7.3. Detailed sedimentary log of the study area with units and facies codes.

... 139

Figure 7.4. Photographs demonstrating the facies present (A) sand volcanoes with

sandstone sheet comprising facies C1. (B) View facing NE, of units 4, 5 and 6. Yellow notebook sits on unit 4. Red arrow denotes thickness of facies A1. (C) Photomontage of units 1 and 2 showing deformation in facies B1. (D) View

(19)

perpendicular to (E) to demonstrate large-scale dewatering structures and lack of directional indicators within this facies. Red arrow denotes extent of facies A1. (E) Bedding plane view of top of unit 3, showing mud diapir (B2) within B1. Person is standing above one edge of diapir. ... 140

Figure 7.5. 3D schematic model of Freagh Point idealised along the plane of a major

growth fault which cross-cuts the outcrop, demonstrating the relationships between and within the facies present. Facies labels and stratigraphic units are shown on the left of the figure. ... 141

Figure 7.6. Photomontage of facies C1 where it is cross-cut by a major growth fault.

Continuous laminations within the sand-sheet are well preserved and have been interpreted (white lines) for clarity. Images flow from top-left to bottom-right. ... 146

Figure 7.7. Scatter diagram of flank angles from representative sand volcanoes. The

data demonstrates the decrease in sand volcano flank angles with increased distance from vent, up to 30 cm. A minor increase in flank angle is seen at 40 cm, or equidistant between vents. ... 147

Figure 7.8. Cross-sectional view through the extruded sandstone sheet highlighting

a sand volcano and positions of back-scattered electron diffraction images of the sandstone sheet demonstrating the decrease in grain-size away from vents. Quartz grains are black. Lens-cap for scale in upper image. ... 148

Figure 7.9. Photomontage demonstrating the relationship between facies C2

(individual sand volcanoes outlined in red) and A1 (white laminations). Scale bar (yellow) 1 m. ... 149

Figure 7.10. Process-based model of extrudite formation. (A) Where multiple,

closely-spaced vents are present, the interaction of gravity currents will produce a sand-sheet with continuous laminations between vents. (B) A single vent will produce a discrete feature, such as a mound due to radial gravity flows. Long term extrusion events will result in channelling, redirecting the radial gravity flows, resulting in deposition away from the extrusion-site as shown to the right of the cartoon. ... 156

(20)

Figure 8.2. Geological Map of the study area with pipes. All pipes occur within the

Carmel Formation or Gunsight Butte Member of the Entrada Sandstone. Modified after Doelling et al. (2000). ... 165

Figure 8.12. Model depicting injection complex architecture and processes. (A) Fluid

Flow from Navajo sandstone. (B) Fluidisation and pipe intrusion with preferential exploitation of conglomeratic channels in Carmel Formation. (C) Laminated Rind facies with placer-style accumulation of heavy minerals in concentric laminations. (D) Hydro-fractured clast breccia in silty facies of the Gunsight Butte Member. (E) Normal and reverse grading in sills. (F) - Extrusion (max. pressure gradient). (G) Exotic lithologies suspended in pipes (Page Sandstone, Carmel Formation, chert pebbles from J-2 unconformity). (H) Fluidisation zones possibly representing vents at the palaeosurface. ... 186

Figure 9.1 Synthesis of the key similarities and differences between

mudstone-hosted, sandstone-hosted and shallow sandstone intrusions based on experimental and field observations combined with those from the literature. ... 213

Figure 9.2 Tripartite architecture of mudstone-hosted intrusions (left, from Hurst et

al., 2011) and bipartite architecture of sandstone-hosted intrusions... 211

Figure 9.3 Summary of the relationship between host strata and the architecture of

sandstone intrusions in the Panoche Giant Injectite Complex. The “sill zone” is marked by the red box. From Hurst et al. (2011) based on Vigorito and Hurst (2010)...214

Figure 9.4. Intrusion margins at a range of scales. Curvilinear intrusion margins

with thick, fine-grained lining. From experiments in Chapter 7. B. Interpreted sketch of a sandstone sill, host rock is white and margin is curvilinear. Flow is from left to right in image. Adapted from Kawakami and Kawamura (2002). C. Undulous margin between injected banded sandstone (SC) and host mudstone (HC) interpreted as a scallop. Adapted from Scott et al., (2009)... 220

(21)

List of Tables

Table 4.1. Values of the dimensionless parameter n as a function of the terminal

Reynolds number. From Richardson and Zaki (1954) and Di Felice and Kehlenbeck

(2000). ...32

Table 5.1. Summary of parameters used in the experiments. Glass bead and silica sand densities were 2.6 g cm-3_{. Hydraulic head was 105 cm in all experiments. Water} velocities were assumed to be approximately 0.3 cm sec-1_{, the lower detection limit} of the flow meter. All experiments had a 1 cm thick middle layer, composed of fine Ballotini (0-44 μm) with <10% Poly V®_{(1.1-1.2 g cm}-3_{) as a tracer. See Figure 5.1 for} details of inlet configuration. ...45

Table 5.2. Wall effect numbers for this and previously published studies... 74

Table 6.1. Calculated volumetric flow rates for each experimental run ...99

Table 6.2. Particle flow velocities at a range of pints across the pipe diameter...106

Table 6.3. Calculated pseudo-fluid density ( and viscosities ( ) from experimental observations...116

Table 6.4. Calculated maximum and minimum flow Reynolds numbers for measured velocities in an active fluidisation pipe (experiment 3) based on psuedofluid values...116

Table 6.5. Calculated flow Reynolds numbers with density and dynamic viscosity values for pure water. ...117

Table 8.1. Facies present within the injections. ... 170

Table 8.2. Mean porosities for each facies averaged over 5 samples. Porosity calculated by point counting...181

Table 8.3. Minimum settling velocities for a clast diameter of 0.7 m as a function of grain concentration in the pseudo-fluid, for the example in Scott et al. (2009)…….198

Table 8.4. Minimum settling velocities for a clast diameter of 3 m through a pseudo-fluid of varying particle concentrations, for a pipe in the Kodachrome Basin. …...199

Table 8.5. Recalculated minimum flow velocities and flow Reynolds numbers for the data in Scott et al. (2009) showing that values have dropped by up to 2 orders of magnitude. ...200

Table 8.6. Calculated Reynolds numbers (Re) for grain concentrations of ɸ = 0.54, ɸ = 0.35 and ɸ = 0.15 for a pipe in the Kodachrome Basin...202

(22)

Nomenclature

A Aperture (m)

Ar Archimedes number

C Cohesive strength

CD Drag coefficient for a particle in relative motion with a fluid

(non-dimensional)

CD, O Drag coefficient for a solitary particle in relative motion with an

infinite fluid (non-dimensional)

d Clast diameter (m)

ds the diameter of a sphere with the same volume as the particle

dc the diameter of a circle of the same area as the projected profile of the particle in its most stable orientation

ε Voidage

e Void ratio

Epot Potential energy

Fk Drag force

Porosity

G Specific gravity

g Acceleration due to gravity H(c) Pipe height / 3

ic Critical hydraulic gradient k Volumetric shape factor

n an exponent, a function of particle shape

Solid volume fraction

Psuedofluid in contact with the larger particle phase Density (kg m-3₎

(23)

Ψ Sphericity / grain shape Normal stress

Internal friction coefficient

Shear stress

Tensile strength

µ Kinematic viscosity (Pa s) U Velocity (ms-1₎

Vd Volume of the injected sand grains

Subscripts

s,p solid f fluid pf pseudo fluid q quartz L larger particle S smaller particle i single sphere mf minimum fluidisation

(24)

(25)

1. Thesis context, significance and structure

1.1. Thesis rationale and objectives

The recognition of sandstone intrusions as an important component in hydrocarbon plays within the last 18 years has promoted a period of intensive research focussed on the subsurface remobilisation of sediment. (e.g. Dixon et al., 1995; Molyneux et al., 2002; Huuse et al., 2004; Hurst et al., 2005; Lonergan et al., 2007) Prior to their recognition in hydrocarbon plays, sandstone intrusions have been recognised in a range of geological settings for nearly 200 years (Murchison, 1827) and ancient examples of extrusions were described in the mid 20th_{century (Gill and Kuenen,}

1957). Although it is widely accepted that sand injection results from the fluidised flow of sand into fractures, driven by overpressured fluids, there remains a lack of process understanding regarding the nature of the flow during active injections and direct observations of the impact on surrounding sediments. Transport and subsequent sedimentation of grains from parent sand bodies has, until now, remained poorly understood and the dynamics of the intrusion process itself have been largely ignored. Given the inability of field and core-based studies to shed light upon the sedimentary processes in active injection, physical modelling offers considerable potential for constraining the flow conditions during injection. As sand injection occurs in a wide variety of sedimentological settings and at a range of scales (e.g. Saucier, 1989; Purvis et al., 2002; Strachan, 2002; Davies et al., 2003; Moreau et al., 2011), it may be assumed that the processes remain largely similar throughout. The modelling component of this project sets out to recreate sand injections in a laboratory environment, building upon previous conceptual ideas, with the aim of understanding the dynamics of clastic injection, from process to product.

(26)

The field-based aspect of this project is framed by a gap in the knowledge surrounding both the dynamics of sand extrusion in subaqueous settings and the nature of injection in coarse-grained successions. Extruded sandstones are known to form stratigraphic traps for hydrocarbon accumulations as sheets or mounds with 4-way dip closure (Hurst et al., 2006; Andresen et al., 2009). As oilfields such as the Eocene Chestnut reservoir in the North Sea may already be producing from extruded sandstones (Huuse et al., 2005), there exists a need to further the understanding of their formation dynamics and geometries in order to aid their recognition in the future.

Although many of the studies in the recent surge of literature have been focussed on the nature of injections in deep-water, or at least, fine-grained successions, there remained a lack of investigations into injections in continental strata. Aeolian plays can comprise vast reservoirs such as the Argyll/Ardmore filed in the Central North Sea (Gluyas, 2005) and injections have been recognised in such successions in outcrop (Netoff, 2002; Chan et al., 2007; Loope et al., 2013; Rowe, 2013). It must be considered however, that such intrusion will be practically invisible in 3-D seismic data and the identification of facies may prove enigmatic in core. Therefore a systematic investigation of the geometries and facies of sandstone injections hosted in the aeolian Entrada Sandstone of the Colorado Plateau is undertaken with the aim of highlighting their significance in coarse-grained strata and shedding light on the complex relationships between injection and host strata from a new perspective.

1.2. Thesis structure

This thesis begins with an introduction to the concept of fluidisation and the conditions required for inducing sand injection. This is followed by a theoretical review of fluidisation which discusses relevant literature from a chemical engineering point of view to support the inferences made in later chapters. This leads into a review of fluidisation in geological systems in the form of sand

(27)

injection. Results in the form of four independent research chapters are presented, each with their own rationale, discussion and conclusions, two of which are published, one is in review and the fourth is in preparation for publication. The thesis is concluded with a synthesis of the key advances made in relation to the existing knowledge of mudstone-hosted sandstone intrusions.

(28)

2. Outcrop and seismic based studies of

sandstone intrusions

2.1. Fluidisation inferred from geological systems

Fluidisation is known to be important for the formation of a wide range of geological features, including: i) en masse dewatering structures, ii) clastic dykes and sills, and, iii) sand volcanoes, sand sheets and other extrudites (e.g., Maltman, 1994; Jolly and Lonergan, 2002; van Rensbergen et al., 2003; Gallo and Woods, 2004; Hurst et al., 2006; Vigorito et al., 2008; Rodrigues et al., 2009). Sandstone intrusions are the product of the remobilisation of unconsolidated sediments in the subsurface by the flow of fluids, most often basinal waters produced during consolidation and de-watering of underlying sediments (Jonk et al., 2005). At low fluid velocities, pore fluid is able to percolate through sediment without creating a local difference in velocities between fluid and sediment grains (liquefaction) or suspending the sediment (fluidisation), this is termed seepage (Lowe, 1975).

The most detailed description of classical fluidisation in nature from the so-called boiling sand springs in Nebraska (Guhman and Pederson, 1992) where the upwards flow of groundwater through cylindrical conduits acts to suspend sand grains, although no net-flow of sand occurs. Guhman and Pederson (1992) report this phenomenon in a series of springs close to the Dismal River and attribute the localised flow of water to secondary permeability in buried cohesive units. Springs are up to 10 m in diameter and have been plumbed to depths of 44 m revealing circular cross-sections and defined walls. The “boiling” sand layer consists of a mixture of sand grains and water, where the sand grains are held in suspension and are overlain by clear water, with a distinct interface separating them. In this case the

(29)

sand is fluidised by the upward-flowing water, and as the authors report a “churning action” in the suspended sand, it possible that the fluidised column is exhibiting heterogeneous behaviour (Section 3.3). During measurement of spring depth using a 9 kg weight on the end of a line, sharp tugging was felt at 26 m depth, this could possibly be due to “necking” of the conduit causing increased turbulence (Guhman and Pederson, 1992).

2.2. Elements of a sandstone intrusion complex

A comprehensive review by Hurst et al., (2011) has provided an elegant classification of the elements of an intrusive complex and these are summarised in Figure 2.2. An intrusion complex can generally be subdivided in four components; (i) a parent sand body; (ii) dykes; (iii) sills; and, (iv) extrusions. Sandstone sills are approximately concordant with the fabric of the host strata and are generally considered to be tabular bodies but are known to exhibit ”stepping” with minor discordance connecting adjacent sills. Dykes are generally discordant and cross-cut stratigraphy and are termed low-angle dykes up to 20° and high-angle above 20°. Although the review of Hurst et al., (2011) is inclusive of most aspects of sand injection, owing to bias in the pre-existing literature towards injection in mudstone-dominated successions, much of the review therein refers to injections in deep-marine environments. This section aims to briefly appraise the current understanding and introduce additional elements that have previously been overlooked due to their small scale or recent identification.

(30)

Figure 2.2. Schematic block diagram showing the elements of a sandstone injection

complex. Parent sand body geometry is marked by dashed line. Modified from Hurst et al. (2011).

2.2.1. Parent beds

A non-exhaustive list of parent sand-bodies for injection complexes has been complied by Hurst et al., (2011) with 11 out of the 15 studies based on deep-marine depositional elements (Dixon et al., 1995; Surlyk and Noe-Nygaard, 2001; Hillier and Cosgrove, 2002), Purvis et al., 2002; Strachan, 2002; Duranti and Hurst, 2004; Briedis et al., 2007; Hamberg et al., 2007; Hubbard et al., 2007; Lonergan et al., 2007; Satur and Hurst, 2007) and two studies being based on shallow-marine systems (Obermeier, 1996; Hildebrandt and Egenhoff, 2007), demonstrating the bias in recognition from deep-marine successions. This may in part be due to differential erosion of mudstone and sandstone enabling their enhanced recognition in outcrop,

(31)

but is primarily due to the density difference between mudstone and clearly discordant sandstone bodies which is so easily discernible in seismic data (Fig. 2.3).

In all cases however, the parent unit shows extensive dewatering in the form of; (i) dish and pillar structures (e.g. Surlyk and Noe-Nygaard, 2001; Duranti and Hurst, 2004), (ii) inclined laminae (e.g. Dixon et al., 1995); Purvis et al., 1992) and/or, (iii), a degree of homogenisation and loss of depositional fabric (e.g. Obermeier, 1996; Hamberg et al., 2007). As suggested in Figure 2.1, the external geometry of a parent unit is often heavily modified by remobilisation and subsequent injection and is often transformed from lenticular channels to lenses with mounded tops (e.g. Duranti and Hurst, 2004; Wild and Briedis, 2010) and “wing-like” margins (e.g. Huuse et al., 2007; Jackson, 2007; Cartwright, 2010). Many of these examples are limited to remobilisation at depth (> 100 m) and little research has been conducted into the effects of fluidisation on shallow-buried sandstone bodies with the exception of Hildebrandt and Egenhoff (2007) and Oliveira et al., (2009). The model of Hildebrandt and Egenhoff (2007) is scrutinised in detail in section 7.6 with regard to the proposed seafloor extrusion above shallow-marine massive sands, however, the internal structures reported from shallowly buried remobilised bodies are consistent with their deep-water counterparts with pervasive dish structures, flame structures and structureless zones (Hildebrandt and Egenhoff, 2007; Oliveira et al., 2009; 2011)

(32)

Figure 2.3. Conical sandstone intrusions from the Faroe-Shetland basin showing

characteristic discordance. Arrows denote the tops of the intrusions. From Cartwright et al. (2008).

2.2.2. Dykes and Sills

Sandstone dykes and sills form the intrusive element of mudstone-hosted injection complexes and their dimensions, external geometries and internal sedimentological textures have been studied extensively in core and outcrop (e.g. Peterson, 1968; Taylor, 1982; Duranti and Hurst, 2004; Diggs, 2007; Parize et al., 2007; Thompson et al., 2007; Scott et al., 2009; Sherry et al., 2012). Tabular or saucer-shaped sills can extend laterally for kilometres and dykes can cross-cut over 100 m of stratigraphy (Huuse et al., 2005b; Vigorito et al., 2008). It is thought that the macro-scale (m to km) geometry of these intrusive elements is controlled primarily by the process of hydrofracturing (Levi et al., 2011) with the immediate (< 1m ) relationship between the intrusion and host strata being dominated by erosive processes such as corrasion and abrasion (e.g. Kawakami and Kawamura, 2002: Diggs, 2007; Scott et al., 2009; Kane, 2010). Erosion is not observed along the entire margins of sandstone injections and much of the eroded material consists of clasts of mudstone or host-rock which is then suspended in the sandstone intrusion (e.g. Scott et al., 2009). The nature of injectite margins and how they relate to flow conditions during

(33)

emplacement is discussed in more detail in Chapter 6. Very little is known about the grain-scale processes that are involved in the incorporation of host strata into the active injection; corrasion is believed to be a dominant process along with hydro-fracturing and subsequent spalling of the host (Scott et al., 2009). Evidence providing the basis of the corrasion theory includes sand grains (from the injecting slurry) being embedded in (host) mudstone clasts (Scott et al., 2009), however, there is no way to discern at what point in the intrusion process the grains became imbedded, thus highlighting the need for further experimental study of injection processes.

Dykes are not always linear features and geometries include; tapering (Strachan, 2002); bulbous and curved (Parize et al., 2007); bifurcating (Hubbard et al., 2007); and planar features (Duranti and Hurst, 2004). Outcrop and core-based studies are inherently limited in terms of spatial analysis unless comprehensive 3-D exposure of an injection complex in outcrop occurs. Sills are generally tabular features but have been reported as showing bifurcation (Truswell, 1972) and stepping (Parize et al., 2007) although their geometry is likely to be inherently linked to the stratigraphic architecture of the horizontal strata into which they intrude.

2.2.3. Sandstone extrusions

Extrudites or sandstone extrusions are bodies of sand that have been vented following the intersection of a sand injection with the surface and have been described from a wide-variety of stratigraphic settings and on a range of scales from centimetres to kilometres (Chapter 7.1). Often strongly associated with seismic activity (Obermeier, 1996; Quigley et al., 2013), extrusions can take the form of sand volcanoes, such as the well-known examples up to 1.5 m in diameter from the Carboniferous of County Clare, Ireland (Gill and Kuenen, 1957; Strachan, 2002) but can also produce extruded “sheets” of sand covering up to 250 km3_{, as in the North}

Sea subsurface (Løseth et al., 2012). These large-scale examples are described from seismic data and the extruded sand purportedly displays features such a wedging-out radially and gently dipping internal reflectors (Løseth et al., 2012), similar

(34)

characteristics to outcrop-observed counterparts (Hurst et al., 2006). Analogue experiments have been conducted which investigate the distribution of fluidised sediment once erupted through a fracture (Rodrigues et al., 2009; Ross et al., 2011). The experimental observations of Rodrigues et al. (2009), although heavily cited by Løseth et al. (2012), could be called into question as air is utilized therein as the ambient medium into which sediment was ejected. The implications of this ambiguity and the dynamics of extruded material and how extrudites are formed are discussed in detail in Chapter 7.

2.3. Injection in coarse grained environments

Studies investigating the remobilisation of sand and subsequent injection into a coarse-grained host remain somewhat behind the level of research into mudstone-hosted injections. The numerous examples of sand injected into mud in the literature highlight their increasing recognition as a key component of deep-water clastic systems (e.g. Parize and Friès 2003; Huuse and Mickelson 2004; Jackson, 2007; Shoulders et al., 2007; Vigorito et al., 2008; Cartwright 2010). However it is clear that remobilisation and injection do occur in coarse-grained clastic systems, and on a variety of scale, ranging from cm-scale dykes and sills (Glennie and Hurst, 2007; Hurst and Glennie, 2008) to decimetre-scale “mega-pipes” and columnar intrusions (Hannum, 1979; Netoff, 2002; Huuse et al., 2005b; Chan et al., 2007). As the fine-grained sealing mudstone inherent to deep-water systems is mostly absent in these coarse-grained settings, other strata such as evaporites and carbonates with low vertical permeabilities (Chan et al., 2007) can prevent early dewatering and subsequent overpressuring of pore fluid in the parent bed. The aim of Chapter 8 is to investigate how the injection processes in aeolian strata, may differ from those in fine-grained hosts in terms of initiation and development. Chapters 5 and 6 detail experimental modelling of sand injection in coarse-grained strata.

(35)

3. Clastic injection: liquefaction,

fluidisation and injection processes

3.1. Introduction

Soft-sediment deformation (ssd) is considered to occur in unlithified sediments or sedimentary rocks that are not entirely lithified as a result of intergranular movement in response to an applied stress. A combination of forces acting on an unstable sediment is often responsible for the genesis of ssd structures, the most ubiquitous being gravity, which results in Rayleigh-Taylor instabilities, downslope-movement and compaction (Maltman, 1994). If sediment has not undergone sufficient compaction and remains underconsolidated with high porosity, low-cohesion and inter-granular bonds, it will be susceptible to remobilisation. This requires grains to be mobilised by the fluid flow through the sediment (the most prevalent source of fluid being water lost from consolidation of surrounding sediments) and subsequent upwards transport of those grains to a new locus as an injection. Subsurface fluid migration is driven by a pressure differential between the pore fluid and either a shallow aquifer or the surface/seabed.

3.2. Seepage & liquefaction

Grains of non-cohesive sediment are acted upon by three forces; (i) the tractive force of the cumulative flow driven for example by the hydraulic head; (ii) the local seepage force, and; (iii) gravity. If the flow of fluid through the sediment exerts no net force upon the grains and results in little to no grain reorganisation, it can be termed seepage. The phenomenon of liquefaction occurs when an external factor, such as cyclic loading from seismic shock pressurises pore fluid in sediment and destroys a metastable particle configuration (Hird and Hassona, 1990). If pore

(36)

pressure is increased to equal that of the burial pressure, the effective stress on the sediment / pore water mixture is therefore zero and the sediment acts as a viscous fluid, showing no shear strength (Maltman, 1994). Rounded grains have been shown to liquefy more easily than angular grains or sediments containing significant amounts of platy minerals, which increase the compressibility and cohesion of sediment (Hird and Hassona, 1990).

Yamamuro and Lade (1999) tested natural sands for their liquefaction potential using undrained triaxial compression testing. This technique determines the shear strength of a material by the deviation from the measured compressive strength. They show that most earthquake-induced liquefaction occurs in loose, silt rich sands, although most laboratory studies on liquefaction have been performed on clean sand and that there is a strong correlation between the liquefaction potential of a soil and its fines content (Yamamuro and Lade 1999). Isotropic compression was conducted to create confining pressures (25 – 500 k Pa) and shear tests were commenced immediately before creep could occur in the test sample. At low confining pressures (< 125 k Pa), complete static liquefaction occurs whereas at higher confining pressures the sample underwent temporary liquefaction and showed dilatant behaviour. Liquefaction potential was also tested by increasing the fines content, from zero to 50% and it was found that liquefaction potential increased with increased fines content. When compression tests with no fines were carried out, the sample did not undergo static liquefaction. Yamamuro and Lade (1999) suggest that increasing the percentage of fine grained material in a sample results in the effective stress paths being depressed, with smaller differences in initial peak stress. The authors also studied the axial strain in the sample and noted that it decreases as the fines content is increased. They concluded that despite the increase in sample density through increased fines content, the liquefaction potential increased and therefore neither void ratio, or density of sand should be used as an indicator of liquefaction potential. These findings are confirmed by Shapiro and Yamamuro (2003) who showed through laboratory testing that the

(37)

presence of silt greatly increased the compressibility of otherwise clean sand. In order to explain this behaviour, Yamamuro and Lade (1999) put forward a hypothesis regarding the structure of sediments with liquefaction potential; if silty-sand is deposited under low-energy conditions the particle-structure is highly compressible. The fabric has a loose structure with fine grains occupying the void space between larger particles. Most of the load placed on test sediment with this type of structure is supported by the larger grains, with fine grains merely acting to increase the overall density of the sediment, without affecting overall behaviour. Nevertheless, some of the fine grains will be present at the contacts of the larger grains, sometimes holding them apart; it is these particle interactions that are responsible for the compressive behaviour of the sediment. When the system is compressed and sheared, these fine-particles are forced from between the grain contacts and into the void spaces, thus decreasing the overall volume of the sediment (Figure 3.1.)

Figure 3.1. Compressibility of mixed particle systems; (a) as deposited; (b) effects of

compression and shearing. Redrawn from Yamamuro and Lade (1999).

Experimental triaxial testing of silty sediments found that increasing fines content from 0-44% decreased the liquefaction resistance of fine to medium sand, with this trend being revered for fine fractions above 44% (Xenaki and Athanasopoulos, 2003). This is known as the transition fines content, and up to this fines content,

(38)

compressive behaviour is mainly controlled by the larger grains. Beyond this critical threshold, compression behaviour is controlled by the fine grains. This threshold can be anywhere between 20% and 44% depending on the fine fraction, volume fraction and stress conditions (Monkul and Ozden, 2007). It was found that in mixed sediment, shear strength decreases with a corresponding increase in silt content whereas for a layered sample with the same amount of silt, shear strength in unaffected (Naeini and Baziar, 2004). If the mean grain diameter ratio (D50-sand/d 50-silt) also influences liquefaction potential with small ratio resulting in increasing

liquefaction potential corresponding with increasing fines content. As D50-sand/d50-silt

increases, liquefaction potential of the silty sand becomes close to that of clean sand (Monkul and Yamamuro, 2011)

Under undrained conditions, when pore-water is still present in a sample, pore pressure will increase and the potential for liquefaction becomes obvious (Yamamuro and Lade 1999); this situation is most applicable to sediments at the surface where compaction has not occurred. It is this uncompacted structure proposed by Yamamuro and Lade (1999) that may also be responsible for increased dilatancy with increased confining pressures, as well as decreasing compressibility of the sediment and increasing the “stiffness” of the sediment structure.

3.3. Transition from liquefaction to fluidisation

Seepage is unlikely to create sedimentary structures unless the system becomes liquefied, by seismic shock for example, or an increase in fluid velocity fluidises the system. According to Lowe (1975), the term seepage can also be applied when pore-fluid flow rates are above the minimum pore-fluidisation velocity, but the sediment is either highly compact or confined by an overlying permeable layer as drag on the sediment particles is negated by compaction. This scenario can be represented graphically (Figure 2.1) where the area below the dashed line represents a fixed bed of particles, where contacts between the particles negate the difference between

(39)

weight and particle-fluid interaction forces. Between the two curves, particles are suspended and fully supported by the fluid. Above the solid curve, drag force exceeds particle weight, but only in a system with an overlying confining layer, restraining particle movement (Di Felice, 2010). This would be termed seepage by the classification of Lowe (1975).

Figure 2.1. Minimum fluidisation and terminal Reynolds number as a function of

Archimedes number. The area above terminal condition denotes seepage at high fluid velocities where the drag force exceeds particle weight. Adapted from Di Felice (2010).

The upper, “seepage” zone is probably very common in natural geological systems, but most fluidisation features form when the confining overlying layer is impermeable and pore pressure builds up in the saturated bed. In order to fluidise grains in a sealed system, there must be a differential pressure gradient across the

(40)

bed (Jolly and Lonergan, 2002). If a saturated, overpressurised bed is capped by an impermeable unit, the overburden pressure is partially supported by the pore fluid, and partially by grain-grain contacts. Overpressure occurs when the pore fluid pressure in a sealed bed is greater than the surrounding, or hydrostatic pressure (Maltman, 1994). Sudden hydraulic fracture of the sealing unit occurs when overpressure in the sealed bed reaches a critical threshold and would create the required pressure gradient, fluidising the grains and entraining them. Once excess pressure has been released, fracture propagation stops, and the grain-water mixture freezes as a clastic intrusion or sand injection (Jolly and Lonergan, 2002). If the overlying strata are cohesive, fracture occurs as planar dykes orientated parallel to the least compressive strength (Jolly and Lonergan, 2002) and where overlying sediments are non-cohesive, intrusions occur as pipes (Nichols, 1995, Ross et al.¸ 2011). Flow regimes in fluidisation pipes have yet to be determined although theoretical estimates have been calculated based upon grain-size and textural relationships in sand injections (Duranti and Hurst, 2004; Scott et al., 2009; Sherry et al., 2012). It has been suggested that flow in liquefied systems is dominantly laminar and fluidisation is characterised by a turbulent flow regime (Duranti and Hurst, 2004; Hurst et al., 2011) although Di Felice (2010) clarifies this as systems with Ar < 10 are dominated by viscous forces and are therefore laminar, and those with Ar > 105 _{are dominated by inertial forces and are therefore turbulent.}

3.4. Fluidisation

Once sediment has been liquefied it is primed for fluidisation and will remain in this state until pore pressure is reduced below the burial load (Maltman, 1994). Fluidised sediment injection can only occur if fluid migration through the sediment driven by a pressure gradient is of a sufficient velocity to overcome any resistance the grains have to mobilise due to friction thus entraining the grains as a slurry which continues to flow until the pressure gradient has been equalised.

(41)

3.5. Injection processes

3.5.1. Overpressure

In order to initiate fluidisation and sand injection, a parent bed must be overpressured, in that pore-fluid pressure must be elevated, and will build as a function of the sealing capacity of the host and the rate of fluid charge into the parent bed (Hurst et al., 2011). The change in the sealing capacity of the host rock can be related to the mechanical compaction rate of mudstone, or diagenesis and subsequent cementation of sandstone. The development of overpressure can also be driven by the reduction in pore volume within the parent bed due to disequilibrium compaction and may also be driven by tectonic compression on a basinal scale, a relatively rapid method of building overpressure (Osborne and Swarbrick, 1997). On a smaller scale, overpressure can be driven by a hydraulic head difference over a river channel levee for example (Li et al., 1996; Singh et al., 2001) or due to artesian flow in an aquifer (Dragantis and Janda, 2003) provided a sealing lithology is present above the aquifer. Diagenetic phase transitions such as the opal A to CT can release water within a buried parent sand body (Davies et al., 2006), and provided a seal is present, can produce overpressure and prime the reservoir for fluidisation when triggered.

3.5.2. Triggering

It is the catastrophic release of overpressure in the parent sand body and fracturing of the sealing lithology that ultimately results in sand injection and the mechanism that triggers this release is varied. The most cited cause of fluidisation and sand injection is seismic activity (e.g. Saucier, 1989; Obermeier, 1996; Moretti, 2000; Boehm and Moore, 2002; González de Vallejo et al., 2005; Huuse et al., 2005b; Hildebrandt and Egenhoff, 2007; Moretti and Sabato, 2007; Alfaro et al., 2010) due to the cyclic loading imparted into the overpressured reservoir, although until recently, evidence for earthquake induced sand injection was only available from the shallow subsurface (e.g. Saucier 1989; Obermeier, 1996). Cyclic stresses can also